Clinical experience with immunomodulatory therapies in patients with sepsis has been disappointing to date. An important example is the clinical experience thus far with tumor necrosis factor soluble receptor (TNFsr). In early preclinical work utilizing endotoxin or bacteria infusion challenges, TNFsr appeared protective. However, in clinical trials, this agent appeared harmful in patients with sepsis. These findings have emphasized that patients with sepsis may not he uniform, and important factors that differ between patients may influence the fundamental effects of immunomodulatory therapies. Such factors include site and severity of infection as well as dose of immunomodulatory agent. In order to study the influence of such factors on the effects of immunomodulatory therapies with sepsis, we have developed an experimental system using a small animal model and a multifactorial design to evaluate the effects of site and severity of infection and dose of immunomodulatory therapy. Using this model, we plan to study the differential effects of placebo, low or high dose TNFsr during mild or severe bacterial intravenous, intraperitoneal or intrabronchial infection.